Archive for May, 2014

My company Aditi did a candid part tongue-in-cheek piece on me recently. No good techie facts but if you want to know about the geek from which all this information spews take a read.

And yes… I really am BAT-MAN!!

This is the first module of substance in the my new Microsoft Azure Administration mini-series hosted primarily off the Aditi Technologies blog that parallel my just-release Pluralsight course entitled “Microsoft Azure Administration New Features (March 2014)” .

There are two main components to the Recovery Services functionality. The first is called Backup Recovery Services and is used to do automated backups of data primarily from on premises into the Azure Cloud. The advantage from a security and Disaster Recovery standpoint is that the backups are stored off site from your data center in the safe Azure Cloud storage. Recall that Azure storage gives you automatic 3x free replication within that same data center as well. Before the data is transmitted it is encrypted as it is when it is stored in azure storage. Backups are done incrementally to allow point—in-time recovery. It does incremental backups also to improve efficiency, minimizes transfer time, and reduces storage costs.

The second main part is of Azure Recovery Services is the Hyper-V recovery manager (HVRM). This works in conjunction with System Center Virtual Machine Manager to backup Virtual Machines to the Azure Cloud. So why would you be interested in the Hyper-V Recovery Manager Azure feature? The Hyper-V Recovery Manager is used for Disaster Recovery (DR). When you talk about High Availability (HA) you talk about running your app on a single data center so that your app stays running but possibly in a degraded mode or slower response. But with HA your app does not go down. When we refer to DR, there is failure across the VM either in pieces or as a whole where your entire data center goes down. So you need a solution that brings up your app in a secondary data center in a short amount of time so your business is not affected significantly. HVRM will help you manage your DR from one DC to another. HVRM is the go-between to manage this process

HVRM works with Hyper-V Replica in Windows Server 2012. A host running VMs in a DC can replicate all the VMs into another DC. The challenge is most customers have many VMs so you have to orchestrate the order and timing in which VMs come up in the new DC during DR. You could use piece together a PowerShell or System Center Orchestrator that you could put together a DR solution process. But this is sort of complex so HVRM gives a simple solution to this problem for both the VM and the data it uses. As long as the data is in a VM (in a VHD) you can replicate data transparently.

HVRM is in Azure and use the same Azure account to manage your on premises VM. HVRM gives you a custom orchestrated way for recovery. Recovery scripts that normally run in your primary DC can now run in azure. Azure will monitor the two DCs and orchestrate the DR if one of them goes down. You install a HVRM agent on the VMM host machine (not on all the VMs in the group) and pushes metadata to Azure (but your data stays in the DC). The metadata is what you see on the VM console in Azure that is sent to Azure regularly.

Before we can fully understand recovery services there are a few important Azure recovery services concepts to be understood, such as Azure storage vaults, their associated certificates, and the backup agent.

An Azure “vault” is a logical abstraction onto Azure blob storage. When choosing to backup data in the Azure Cloud (blob storage), either through Backup Recovery Services or Hyper V Recovery Manager, you must create a backup vault in the geographic region where you want to store the data. A vault is created using Windows Azure PowerShell or through the Azure portal using a “Quick Create” mechanism.

You do not set the size of an Azure vault. Since a vaults maps to Azure page blob storage, the limit of the entire vault is set at 1 TB, which is the size limit of an Azure page blob. But the limit on actual stored data on an Azure backup vault is capped at 850GB. That’s because with vaults there is metadata associated with the backup, and it consumes around 150 GB of storage if the blob was completely full. Thus, this leaves about 850 GB of actual storage space. You can have more than one server using an azure storage vault. It’s up to you how you want to architect your storage layer.

Vaults required you to register an X.509 v3 certificate with your servers that are using backup vaults. You can obtain one of these certificates by getting a valid SSL certificate issued by a Certificate Authority (CA) that is trusted by Microsoft and whose root certificates are distributed via the Microsoft Root Certificate Program. Or alternatively you can create your own self-signed certificate using the MAKECERT tool. You download the latest Windows SDK to get access to this tool, then run it with a command similar to this one.

This creates a certificate and installs into the Local Computer \ Personal certificate store.
makecert.exe -r -pe -n CN=recoverycert -ss my -sr localmachine -eku -len 2048 “recoverycert.cer”

To upload the certificate to the Windows Azure Management Portal, you must export the public key as a .CER formatted file. And whether you purchase or build your own self-signed certificate you end up with a certificate in .CER file format which mean it does not contain the private key. The certificate must live in the Personal certificate store of your Local Computer with a minimum key length of at least 2048 bits.

When the certificate is installed on the server to be backed up it should contain the private key of the certificate (PFX file), in the case of Hyper-V Recovery Services. So if you will be registering a different server than the one you used to make the certificate, you need to export the .PFX file (that contains the private key), copy it to the other server and import it to that server’s Personal certificate store.

The high-level certificate management steps differ a bit if you are using the certificate for a Backup Vault or a Hyper-V Recovery Vault.
Backup Vault certificate management process:
1. Create or obtain a .CER certificate
2. Upload the .CER file to Windows Azure portal Recovery Services vault

Hyper-V Recovery Vault certificate management process:
1. Create or obtain a .CER certificate
2. Export is as a .PFX file
3. Upload the .CER file to Windows Azure portal Hyper-V Recovery vault
4. Import the .PFX file onto the VMM servers to be backed up

Azure Backup Recovery services requires an agent to be installed on any source machines for a file transfer operation to a backup vault. An agent is a piece of software that runs on the source client machine to manage what is uploaded to the Azure cloud for backup. Note that the Hyper-V recovery Manager does not require an agent since it copies an entire VM’s (its metadata), whereas Backup services can back up as little as one file.

The backup agent is downloaded by connecting to the Azure portal from the server to be backed up. You will go to a specific vault and click on Install Agent under Quick Glance. There are two versions from which to choose which one you will install. The tool you will use to manage the backup will determine which one you choose.
• Windows Server 2012 and System Center 2012 SP1 – Data Protection Manager
• Agent for Windows Server 2012 Essentials
After the installation of the agent is complete you will configure the specific backup policy for that server. To do this you can use the following tools. Again, whatever tool you will use will determine which version of the backup agent you install on that server.
• Microsoft Management Snap-In Console
• System Center Data Protection Manager Console
• Windows Server Essentials Dashboard

If you go into my course on Pluralsight ( you will find a lot more information on Recovery Services plus be able to watch videos of how to use these concepts. You will also learn how to manage Azure Backup Services, see how to take and restore backups, Hyper-V Recovery Managera and its use of Hyper-V vaults. In addition, the course includes a wealth of information and demos on the Azure Scheduler, Azure Traffic Manager, Azure Management Services, Azure BizTalk Services, HDInsight, and improvements to Azure Storage and Web Sites. Hope to see you there!

I just released a new Pluralsight course “Microsoft Azure New Administration Features (March 2014)” on some of Azure’s new administration features which can be found at It covers new features from Azure Q4 2013 leading up to the BUILD release focused on advanced topics for the administration of Microsoft Azure. Topics are primarily focused around Azure IaaS but overlap at times into the Azure PaaS space.

Azure Recovery Services
Backup Vault – Backup on premises data to Azure storage
Hyper-V Backup Recovery Manager – to back up and manage VMs

Azure Scheduler
Developers give IT Ops jobs the write and IT Ops schedules and runs them

Azure Traffic Manager
Deploying PaaS Cloud Services to ATM
Deploying IaaS VMs to ATM

Azure Management Services
Billing and Service Alerts
Auto-Scaling of Azure IaaS VMs
Endpoint monitoring
Operation logs

Azure BizTalk Services
B2B messaging and configure an Azure BizTalk Services Deployment

Azure HDInsight
IT Ops takes the applications created by the developers and configuring it to run and work correctly by deploying to an HD cluster

My company, Aditi, and I are publishing an eight-part series on some of the newer administration features of Microsoft Azure. We will publish weekly off the Aditi site and I will provide a summary and link to it from this blog each week. So if you are following my blog you will get regular updates when it is published. Take a look at the initial overview post describing all that we will discuss over the next eight weeks. If you want more details and demos you can watch the course content (which the blog posts parallel) at the web site. Search for courses with my last name (McKeown) or for the new course entitled “Microsoft Azure Administration New Features”. This course is to be released hopefully sometime this week (week of May 19th).

Over the eight-week mini-series we will cover new features from Azure Q4 2013 leading up to the BUILD release focused on advanced topics for the administration of Microsoft Azure. Topics are primarily focused around Azure IaaS but overlap at times into the Azure PaaS space. From a backup standpoint Azure offers Recovery Services to prevent data loss via Azure Hyper-V Recovery Manager and Recovery Services. Azure scheduler gives you the ability to schedule jobs that can run internal and external to Azure. Traffic Manager supports both PaaS configurations based upon round robin, failover, and performance metrics. Management Services gives you the ability to set up billing alerts, as well as service alerts for services, storage, and web sites and run queries on the results. Virtual Machine Auto-Scaling has been improved to now give you configurable options for both PaaS Cloud Services, Web sites, and IaaS VMs. Azure BizTalk Services integrates the enterprise and the Azure Cloud by configuring B2B messaging. HDInsight is Azure’s solution to bring big data processing to the Azure Cloud. Improvements to Azure Storage include read-only access to geo-replicated storage, improved storage analytics, and allowing the transfer of large amounts of data via physical drive using the Azure Import and Export Service. For Web sites we look at the recent enhancements of AlwaysOn Support, Web Jobs, and Staged Publishing.